We have made significant progress in three major areas related to protein dynamics, folding, and function. (1) Water transport through molecular channels. We performed molecular simulations of the simplest molecular channel, a carbon nanotube, as a model system for transmembrane water transport through pores such as aquaporin-1 (Hummer et al., Nature 414, 188, 2001). We showed that the resulting single-file transport quantitatively follows a random walk (Phys. Rev. Lett., 2002). We also characterized the molecular mechanism of filling and emptying the channel (Waghe et al., J. Chem. Phys., in press, 2002). We showed how small changes in the local polarity result in transitions between filled and empty states (Hummer et al., Nature 414, 188, 2001; Waghe et al, J. Chem. Phys., in press, 2002). We also showed how variations in the electrostatic interactions can drive such transitions (Subramaniam, Rasaiah and Hummer, in preparation). This can explain the functional role of hydrophobic channels in proton pumping proteins such as cytochrome c oxidase. With a membrane setup, we could directly simulate the transmembrane water flow under an osmotic gradient at molecular resolution (Kalra et al, in preparation). (2) Protein and peptide folding. Formation of amino-acid contacts is one of the fundamental steps in the folding of proteins. We have performed microsecond simulations of small peptides in solution, orders of magnitude longer than typical all-atom simulations. This allowed us to compare the simulations directly to the measured loop-closure kinetics (Yeh and Hummer, J. Am. Chem. Soc. 2002). The simulations showed that amino-acid contacts form rapidly, on a time scale of 10 ns, shedding new light on early events in protein folding. (3) Single-molecule atomic-force microscopy is increasingly used to probe rare molecular events such as protein unfolding and ligand unbinding. We have developed a theory that allows us to extract kinetic rates from these measurements, and have applied that theory to estimating rates from the forced unfolding of the muscle-protein titin (Hummer and Szabo, submitted for publication).